The public health burden of falls in older persons is substantial. Several lines of evidence suggest that vitamin D supplements might substantially reduce the risk of falls, potentially by > 25% in persons with low serum 25-hydroxyvitamin D [25(OH)D] levels. However, trial evidence is insufficient to guide policy. The proposed study is a seamless, two stage, Bayesian response-adaptive randomized trial for dose-ranging and efficacy confirmation. The trial is designed to identify the best overall dose of vitamin D supplementation and confirm the level of efficacy of that dose for fall prevention. Participants wil be community-dwelling adults, aged 70+ (~40% black, ~60% women), with a baseline serum 25(OH)D level of 10-29 ng/ml, who report at least one fall in the prior year. After an initial pilt phase to evaluate recruitment strategies and finalize data collection procedures, participants will be randomly assigned to 1 of 4 vitamin D3 (cholecalciferol) test doses: 200IU, 1000IU, 2000IU and 4000IU/d. Participants will take their assigned pills for 2 years with extended follow-up off therapy for 1 more year. Stage 1 of this design will select the best dose of vitamin D for prevention of falls over a 6m follow-up period. Participants in the control (200IU) and best dose group will continue seamlessly into Stage 2 with additional recruitment and continued follow-up of all participants. The principal outcome (Stage 2) is time to first fall (or death) over 2 years f therapy. Secondary outcomes are fall rates, patterns of falling, gait speed, balance, muscle strength, frailty, SPPB score, and, in a sample of participants, 400m walk time and accelerometry. Falls will be ascertained from fall calendars completed daily by participants. Follow-up visits will occur at 3, 12, 24 and 36m. Subgroups with potential for differential benefit from vitamin D supplementation are pre- specified as: a) race (blacks vs others), b) baseline vitamin D levels (10-19 vs 20-29ng/ml), and c) frailty status (frail, pre-frail, not frail). Our prven capacity to recruit a diverse population, including large numbers of blacks, allows us to test the effects of vitamin D supplementation in key subgroups. Strengths of the study include a team of interdisciplinary investigators who have vast experience and an exceptional track record in the design, conduct, analysis and dissemination of clinical trials, many of which have guided policy. The investigative team includes individuals with expertise in vitamin D metabolism, endocrinology, geriatrics, assessment of physical function and falls, biostatistics, epidemiology, central laboratory procedures, adaptive designs, community engagement, and recruitment. Building on our record in previous community-based trials, we expect that trial results will be timely, rigorous, and directly relevant to public health guidelines and will immediately influence fall prevention policies related to vitamin D supplementation in older persons.